Research topics

Data center networking

Virtualization is the corner stone of modern data centers and becomes prevalent also in SMEs. The rationale is simply that virtualization offers a good trade-off between the need to maintain server isolation while sharing (quite) efficiently the ever increasing per server physical resources (CPUs, RAM). In the SigNet group, we aim at exploring the performance of the data path inside and between servers in virtualized data centers. In such an environment, the network starts inside the physical servers due to the need to share the physical network cards between the several virtual machines within the server. This is achieved thanks to a virtual switch. We have ongoing studies on assessing the impact of virtualization from the application (transport layer) perspective.

Data center networks are characterized by a high density of paths and L2/L3 devices. One can also observe a number of key scenario: (i) transfers between VMs within a data center (multi-tier applications up to MapReduce type of workload), (ii) transfer from a server to the end user (e.g., SaaS for residential users up to IaaS for private companies) or (iii) inter data centers transfers (synchronization, VM migration, etc.). As applications delegate to the transport layer and especially to TCP the task of conveying data and of controlling the sending rate, we have carried out some studies on the efficiency of high speed TCP version, esp. TCP Cubic, which is the default TCP flavour in the Linux kernel.

Selected Publications:

Scientific leaders: Guillaume Urvoy-Keller, Dino Lopez, Lucile Sassatelli

Energy and SDN

Energy has become a great concern. Networks themselves are apparently responsible for one third of the total IT consumption (which itself represents 10% of the overall energy consumption). We are tackling this issue along two axes. First, energy consumption in ISPs’ networks and end-to-end energy consumption. Here we focus either on evaluating green protocols or measuring energy consumption due to network activities through experiments or simulations. Second, we currently explore the use of SDN (Software Defined Networks) as a enabler for reducing energy consumption of data center networks through a better management of traffic. This latter topic also relates to the Data center networking theme as we seek to explore the inter-play between network and system virtualizations.

See also the Web page of the project on next generation (SDN based) data centers that describes our research projects Next Generation Data Centers.

Selected Publications:

Scientific leaders: Dino Lopez, Guillaume Urvoy-Keller.

Future generation networks for video content distribution

Video transfer is now the dominant traffic in both residential and mobile access networks. We have started to investigate several issues related to video content distribution. First, in collaboration with Orange Lab (PhD work of Mohamed Bouzian), we aim at modeling the Quality of Experience in mobile access networks. The objective is to enable a network operator to assess the relation between the amount of resources dedicated to serve the users and their actual viewing experience, in particular playback stalls [0]. Within the PhD thesis of Vitalii Poliakov, we focus on routing and caching for video distribution in 5G mobile networks, leveraging network softwarization [0]. We have also started to work on energy- and content-aware planning strategies for convergence of fixed optical and mobile wireless access networks, and have been awarded a UCN@Sophia Labex postdoc grant for two years (starting in Sept. 2016) for our proposal on “Cloud Based Virtualized Optical Networks for 5G Access Network (C-RAN): planning for energy savings and user-level video quality”. Preliminary results have been published.

Selected Publications:

Scientific leaders: Lucile Sassatelli, Ramon Aparicio-Pardo

Mobile Social Networking

The main target of this activity is MANET (Mobile Ad-hoc Networks), where mobile users have to exchange traffic without the help of an infrastructure (no base stations). Owing to node mobility and device heterogeneity, connectivity between nodes is intermittent, leading to volatile networks. To cope with volatility, new internet architectures and protocols have to be developed. Networks that cope with this volatility are named Delay Tolerant (DT) networks. The main challenges of these DT-MANETs are the design of an architecture and protocols that work well in an environment characterized by uncertain networking conditions, high mobility and frequent network partition. Specifically, one has to design transport and routing protocols for DT-MANETs, but also determine the capacity of the DT-MANETs. The major novelty of the approach proposed, besides the inherent cross-layer nature of the problem is the use of network coding. Network coding is a (by now popular in the research community) technique that takes advantage of linear combination of data to lower the global resources needed to transfer data in a network. This technique has proved to be beneficial in specific transport layer design for wireless networks and we want to investigate further its benefits for DT-MANETs.

Selected Publications:

Scientific leader: Lucile Sassatelli

Cognitive Radio

Current wireless networks use the radio spectrum in a very inefficient manner. Indeed, a frequency bandwidth is allocated in a fixed manner to the user, even when this user is not using the radio resource. This type of user is denoted as a “primary” user. When the primary user is not active, a “secondary” user could take advantage of the free spectrum: this is the idea behind cognitive radio. In this activity, the objective is to allow a primary user and a secondary user to transmit simultaneously. Indeed, if the interference caused by the secondary user is somehow known by the primary user, the latter can null out this interference and continue to transmit with the same throughput. To be able to perform this interference cancellation, the different users must share some information (in the extreme case, share all the transmitted messages). The goal of this activity is to explore the cooperation strategies (which information have to be shared) to enable the development of cognitive radio networks where both primary and secondary users transmit simultaneously.

Scientific leader: Luc Deneire

Scientific leader: Jérôme Lebrun